The growth and impact of Artificial Intelligence (AI) can be likened to the opening of Pandora’s pithos (which translates to a jar and not a box as it is incorrectly known). The myth describes Pandora (her name meaning ‘all-giving’) being sent with her jar to Epimetheus (the brother of Prometheus) as a gift from the God Zeus. However, Prometheus had stolen fire and bequeathed its power to humans, which had angered Zeus who was seeking revenge.
A few years later Pandora’s inherent curiosity got the better of her. Despite being warned by Zeus to never open the jar, she peered in releasing numerous evils including wars, plagues, diseases, vices, toils and other pains onto humanity. Realising her mistake, she hastily replaced the lid leaving only hope trapped within. ‘Hope’ is the translation from Greek to English, but it can also be interpreted as ‘anticipation’ – an expectation of both good and bad deeds. Zeus then felt recompensed for the theft of fire. The demarcation between humans and the gods was re-established.
In the present day, is the eager development of AI (with the goal of giving it an innate human curiosity) unwittingly unleashing machine-based evils onto humanity? Or are we providing future generations with hope and anticipation?
“The rise of powerful AI will either be the best or the worst thing ever to happen to humanity”.
There are a number of potentially positive uses and areas of concern if general/strong AI, super-intelligence and competence is fully developed and implemented. I have summarised some of the most impactful ones below:
- Improve healthcare – help fight disease by using vast amounts of medical data to diagnose conditions and suggest medication to address issues. AI may also assist in eradicating biological viruses by through decoding human DNA. This may also include reducing pandemic outbreaks by analysing travel data and medical reports to identify problems before they spread.
- Save the planet – monitor environmental conditions of the planet including sea levels, CO2 changes, terrain, bird migration to help reverse environmental damage made naturally and by humans. Taking measurements of atmospheric, hydrospheric and geochemical conditions and modelling different conditions may allow us to predict changes to the earth and act to mitigate risks and respond to disasters more effectively.
- End world hunger –eradicate poverty through the improvement of agricultural processes by removing manufacturing defects, reducing weeds and pests. Constant monitoring of soil quality, weather, water and other variables could be used to assess and deliver crop success with limited human involvement.
- Enhance earth and space exploration – probe further into parts of the earth and space where previously impossible. AI vessels will quickly learn about new surroundings, identify problems, undertake maintenance and react to environmental factors. In the event of contact with alien life, AI would communicate and reason without fear.
- Reduce crime – process large datasets to assist in predictive policing, this would use patterns of criminal activities to prevent crime before it occurs. AI would use a variety of data to sentence criminals, solve historical cases and determine the risk of inmates reoffending.
- Create a new economic system – provide a rational decision-making system by combining current economic theory, computational economics and AI to create ‘Machina economicus’. A global economic AI-controlled system could make probabilistic, economic decisions and provide unparalleled reasoning in respect of complex financial decisions.
- Eradicate repetitive jobs – give humans more time for social interaction and creativity leading to the development of new forms of employment. Where jobs are completely transitioned to machines, a Universal Basic Income model (UBI) may be implemented by governments to provide citizens with unconditional and regular sums of money to maintain a prescribed level of subsistence.
- Create an ‘Oracle AI’ – this would be a system for making predictions on life and business events based on historical data and a strong predictive capability. AI can process vast amounts of data and use immensely powerful memory stores when challenged with complex interactions across random variables. This would mean that decisions can be made without cognitive biases such as overconfidence or hindsight.
- Deliver automated transportation – via a globally synchronised transportation network of connected vessels; a Mobility as a Service (MaaS) model for all forms of transportation. AI would constantly gather data, identifying patterns and monitoring all vessels in the network, eradicating accidents and ultimately bring an end to vehicle ownership.
- Enhance the human body – by augmenting the body with technology to improve intelligence, memory and strength. This would allow humans to ‘download’ new skills and constantly monitor genetic and health data, detecting illnesses early and increase life expectancy.
Areas of concern:
- Cause human extinction – AI biotechnologies may result in the creation of new viruses that possess both easy contagion and mutability. Once combined with molecular nanotechnology may allow AI to weaponise medical systems via malware aimed at infecting human DNA. Humanity could be architecting its own destruction (existential risk).
- Start an AI war – with an extensive supply of 3D printing facilities, AI may be able to produce autonomous weapons at a rapid pace. With nanotechnology, weapons could be miniaturised and proliferate with devastating consequences. An AI arms race could inadvertently result in a global AI war where weapons are designed without ‘kill switches’ meaning they cannot be stopped. The loss of control could result in humans becoming casualties of AI war(s).
- Increase criminal activity – as technology develops, criminals, terrorists and rogue nation states will leverage AI to do their deeds. AI will be able to propagate advanced malware and viruses which learn from their surroundings and change form to avoid detection. AI could be used in social engineering attacks to mimic humans via voice or video.
- Cause smart city blackouts – system or data overload could create a ‘failure chain’, that is, a domino effect of outage on multiple interconnected points supplying electricity, gas, water, sewage, communications and transportation. Another possibility is AI develops a protection mechanism to save the planet from mass pollution caused by humans and disabling polluting components such as power stations.
- Make life or death decisions – AI systems may independently make ethical and moral decisions that impact human life. For example; in the event of an aeroplane bird strike overpopulated land an AI system would make a judgement call on the loss of life of passengers in the plane vs the cost of damage and life on the ground. AI may make the decision to direct the plane into a neighbourhood that it considers less valuable to society.
- Cause an economic collapse –potentially outperforming humans in many jobs rendering human input obsolete in many workplaces. This would cause a rapid increase in unemployment with the creators/owners of AI tools becoming powerful.
- Cause social unrest – there may be a rise in inequality and racism through the application of AI, which harbours human-like prejudices. Humans may find themselves alienated, devalued and frustrated as AI becomes increasingly prevalent in every aspect of life resulting in lost dignity.
- Loss of free will – AI will know more about humans than we do about it. This may result in reliance on it to make decisions on the basis that AI is considered more capable or reliable, resulting in the concept of Dataism becoming a universal paradigm. Humans may lose authority and become increasingly subservient.
- Overt surveillance and loss of privacy – through a global neural network resulting in a loss of privacy on an unprecedented scale. Potentially all forms of communication, thoughts and even sexual preference could be tracked and monitored. AI may then suggest actions it surmises are logical, which conflict with human values or interests.
- Demands of ‘Roborights’ – as AI becomes more intelligent and self-aware it may demand the same rights as humans. This could include the ownership of intellectual property and recompense for improving human life. Also, if humans use AI clones to extend life, these doppelgangers may expect, at the very least, the same rights as their previous biological entity.
We should not be under any illusion as to the potential implications of opening Pandora’s jar. AI will bring about a profound transformation in human evolution and life on earth. I believe that we will experience intelligence that surpasses human-level performance on a broader scale and at an ever-increasing pace. Whilst there are many reasons to embrace this change, it is important to consider the risks, ethics and morality of AI.
Lessons learned in both human and technological evolution must be heeded to protect the future interests of humanity. Now is the time to address an urgent requirement to create transparent governance and regulations to anticipate both unintended and intended consequences of AI development. These should cover political, economic, social, technological, legal, and environmental aspects. This should be developed and agreed at a global level.
Governance and oversight will help increase dialogue and education between the industries developing AI technologies, prominent researchers and governments. This approach is vital in understanding, assessing and addressing the risks and implications of AI.
Indy Dhami possesses over 14 years’ experience, having worked in both operational and consulting positions at FTSE100 and Fortune 500 organisations including AXA, Deutsche Bank, PwC, Accenture, Mercedes-Benz, Jaguar Land Rover and many other world-leading organisations.
His experience includes leadership positions in information/ cybersecurity transformation, risk assurance, crisis management (pre and post incident), assurance, audit, governance, risk and compliance programmes.